Learning Nouns and Verbs using Cross-situational Statistics

Abstract


 Cross-situational learning is the process of associating words and referents across multiple individually ambiguous contexts. Our experiment uses eye-tracking technology to investigate the effects of word order on the cross-situational learning of nouns and verbs in 2.5 year old children. Participants watch a video in which novel objects and actions are presented simultaneously with novel words, and are tested on their understanding of these words following a five-minute break. The organization of the words is varied across three conditions by modifying the order of the object-referring word (noun) and action-referring word (verb): phrases are presented in noun-verb, verb-noun, or flexible word orders. Our research question investigates whether experience influences learning. As the standard word order in English is noun-verb, we hypothesize that if experience influences learning, participants will learn better in the noun-verb condition. We are also interested in whether nouns and verbs are learned equally well, and whether any differences in learning are related to the word order. Furthermore, children were presented with trials consisting of two familiar nouns and two familiar verbs. Eye tracking data from these trials were compared to test trial data to assess eye movement signatures of noun and verb recognition. Preliminary analyses provide stronger evidence of learning in the noun-verb condition than other word orders. We also see that the speed with which children oriented to the target on noun trials differ from verb trials. We further suggest that the study will contribute to our understanding of language learning and a more reliable interpretation of eye-tracking measures in research.